Automate browser based workflows with AI
Python aiapiautomationbrowserbrowser-automationcomputergptllmplaywrightpowerautomatepuppeteerpythonrpaseleniumvisionworkflow
fix: preserve code_v2 intent after workflow_run DB refresh (#SKY-8261) (#5015)
pedrohsdbbfcd27a5ba6a
diff --git a/skyvern/forge/sdk/workflow/service.py b/skyvern/forge/sdk/workflow/service.pyindex b121a80c6..e9abecab3 100644--- a/skyvern/forge/sdk/workflow/service.py+++ b/skyvern/forge/sdk/workflow/service.py@@ -989,11 +989,19 @@ class WorkflowService:# Check if there's a finally block configuredfinally_block_label = workflow.workflow_definition.finally_block_label+ # Refresh workflow_run from DB to pick up status/failure_reason+ # set by _execute_workflow_blocks. Preserve the original run_with+ # because the DB may now say "agent" for a code_v2 run that fell+ # back to agent (no cached script yet). Downstream code like+ # generate_script_if_needed and _trigger_script_reviewer need the+ # original intent to generate v2 scripts with the :v2 cache key.+ original_run_with = workflow_run.run_withif refreshed_workflow_run := await app.DATABASE.get_workflow_run(workflow_run_id=workflow_run_id,organization_id=organization_id,):workflow_run = refreshed_workflow_run+ workflow_run.run_with = original_run_withpre_finally_status = workflow_run.statuspre_finally_failure_reason = workflow_run.failure_reason@@ -1215,7 +1223,7 @@ class WorkflowService:await self.mark_workflow_run_as_running(workflow_run_id=workflow_run_id, run_with=run_with)# Set script_mode on context so downstream code can skip expensive LLM calls- if run_with == "code":+ if run_with in ("code", "code_v2"):ctx = skyvern_context.current()if ctx:ctx.script_mode = True